Outputs from our so-far best models

Translation

Similar to the training data:

• The Long Prajnaparamita (Toh 8, extract)
• The Diamond Cutter Sutra (Toh 16)
• Heart Sutra (Toh 21)

Somewhat similar to the training data:

• Vinaya (Toh 1-1, extract)
• Madhyamakavatara (Tengyur, Toh 3861)
• Madhyamakalamkarapanjika (Tengyur, Toh 3886)

Very different from the training data:

• Kalacakra Tantra (Toh 362, extract)
• Manjusrinamasamgiti (Toh 360)

Word segmentation and part-of-speech tagging

Coming soon. We will post results here as soon as we set up a test set. We are currently focusing on translation.